1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-24 11:29:10 +00:00
freebsd/bin/sh/TOUR
Jilles Tjoelker bb324af62a sh: Unify EXERROR and EXEXEC
The difference between EXERROR and EXEXEC was that EXEXEC passed along
exitstatus and EXERROR set exitstatus to 2 in the handling code.

By changing the places that raised EXERROR to set exitstatus to 2, the
handling of EXERROR and EXEXEC becomes the same.
2018-11-09 14:58:24 +00:00

302 lines
14 KiB
Plaintext

# @(#)TOUR 8.1 (Berkeley) 5/31/93
# $FreeBSD$
NOTE -- This is the original TOUR paper distributed with ash and
does not represent the current state of the shell. It is provided anyway
since it provides helpful information for how the shell is structured,
but be warned that things have changed -- the current shell is
still under development.
================================================================
A Tour through Ash
Copyright 1989 by Kenneth Almquist.
DIRECTORIES: The subdirectory bltin contains commands which can
be compiled stand-alone. The rest of the source is in the main
ash directory.
SOURCE CODE GENERATORS: Files whose names begin with "mk" are
programs that generate source code. A complete list of these
programs is:
program input files generates
------- ----------- ---------
mkbuiltins builtins.def builtins.h builtins.c
mknodes nodetypes nodes.h nodes.c
mksyntax - syntax.h syntax.c
mktokens - token.h
There are undoubtedly too many of these.
EXCEPTIONS: Code for dealing with exceptions appears in
exceptions.c. The C language doesn't include exception handling,
so I implement it using setjmp and longjmp. The global variable
exception contains the type of exception. EXERROR is raised by
calling error or errorwithstatus. EXINT is an interrupt.
INTERRUPTS: In an interactive shell, an interrupt will cause an
EXINT exception to return to the main command loop. (Exception:
EXINT is not raised if the user traps interrupts using the trap
command.) The INTOFF and INTON macros (defined in exception.h)
provide uninterruptible critical sections. Between the execution
of INTOFF and the execution of INTON, interrupt signals will be
held for later delivery. INTOFF and INTON can be nested.
MEMALLOC.C: Memalloc.c defines versions of malloc and realloc
which call error when there is no memory left. It also defines a
stack oriented memory allocation scheme. Allocating off a stack
is probably more efficient than allocation using malloc, but the
big advantage is that when an exception occurs all we have to do
to free up the memory in use at the time of the exception is to
restore the stack pointer. The stack is implemented using a
linked list of blocks.
STPUTC: If the stack were contiguous, it would be easy to store
strings on the stack without knowing in advance how long the
string was going to be:
p = stackptr;
*p++ = c; /* repeated as many times as needed */
stackptr = p;
The following three macros (defined in memalloc.h) perform these
operations, but grow the stack if you run off the end:
STARTSTACKSTR(p);
STPUTC(c, p); /* repeated as many times as needed */
grabstackstr(p);
We now start a top-down look at the code:
MAIN.C: The main routine performs some initialization, executes
the user's profile if necessary, and calls cmdloop. Cmdloop
repeatedly parses and executes commands.
OPTIONS.C: This file contains the option processing code. It is
called from main to parse the shell arguments when the shell is
invoked, and it also contains the set builtin. The -i and -m op-
tions (the latter turns on job control) require changes in signal
handling. The routines setjobctl (in jobs.c) and setinteractive
(in trap.c) are called to handle changes to these options.
PARSING: The parser code is all in parser.c. A recursive des-
cent parser is used. Syntax tables (generated by mksyntax) are
used to classify characters during lexical analysis. There are
four tables: one for normal use, one for use when inside single
quotes and dollar single quotes, one for use when inside double
quotes and one for use in arithmetic. The tables are machine
dependent because they are indexed by character variables and
the range of a char varies from machine to machine.
PARSE OUTPUT: The output of the parser consists of a tree of
nodes. The various types of nodes are defined in the file node-
types.
Nodes of type NARG are used to represent both words and the con-
tents of here documents. An early version of ash kept the con-
tents of here documents in temporary files, but keeping here do-
cuments in memory typically results in significantly better per-
formance. It would have been nice to make it an option to use
temporary files for here documents, for the benefit of small
machines, but the code to keep track of when to delete the tem-
porary files was complex and I never fixed all the bugs in it.
(AT&T has been maintaining the Bourne shell for more than ten
years, and to the best of my knowledge they still haven't gotten
it to handle temporary files correctly in obscure cases.)
The text field of a NARG structure points to the text of the
word. The text consists of ordinary characters and a number of
special codes defined in parser.h. The special codes are:
CTLVAR Parameter expansion
CTLENDVAR End of parameter expansion
CTLBACKQ Command substitution
CTLBACKQ|CTLQUOTE Command substitution inside double quotes
CTLARI Arithmetic expansion
CTLENDARI End of arithmetic expansion
CTLESC Escape next character
A variable substitution contains the following elements:
CTLVAR type name '=' [ alternative-text CTLENDVAR ]
The type field is a single character specifying the type of sub-
stitution. The possible types are:
VSNORMAL $var
VSMINUS ${var-text}
VSMINUS|VSNUL ${var:-text}
VSPLUS ${var+text}
VSPLUS|VSNUL ${var:+text}
VSQUESTION ${var?text}
VSQUESTION|VSNUL ${var:?text}
VSASSIGN ${var=text}
VSASSIGN|VSNUL ${var:=text}
VSTRIMLEFT ${var#text}
VSTRIMLEFTMAX ${var##text}
VSTRIMRIGHT ${var%text}
VSTRIMRIGHTMAX ${var%%text}
VSLENGTH ${#var}
VSERROR delayed error
In addition, the type field will have the VSQUOTE flag set if the
variable is enclosed in double quotes and the VSLINENO flag if
LINENO is being expanded (the parameter name is the decimal line
number). The parameter's name comes next, terminated by an equals
sign. If the type is not VSNORMAL (including when it is VSLENGTH),
then the text field in the substitution follows, terminated by a
CTLENDVAR byte.
The type VSERROR is used to allow parsing bad substitutions like
${var[7]} and generate an error when they are expanded.
Commands in back quotes are parsed and stored in a linked list.
The locations of these commands in the string are indicated by
CTLBACKQ and CTLBACKQ+CTLQUOTE characters, depending upon whether
the back quotes were enclosed in double quotes.
Arithmetic expansion starts with CTLARI and ends with CTLENDARI.
The character CTLESC escapes the next character, so that in case
any of the CTL characters mentioned above appear in the input,
they can be passed through transparently. CTLESC is also used to
escape '*', '?', '[', and '!' characters which were quoted by the
user and thus should not be used for file name generation.
CTLESC characters have proved to be particularly tricky to get
right. In the case of here documents which are not subject to
variable and command substitution, the parser doesn't insert any
CTLESC characters to begin with (so the contents of the text
field can be written without any processing). Other here docu-
ments, and words which are not subject to file name generation,
have the CTLESC characters removed during the variable and command
substitution phase. Words which are subject to file name
generation have the CTLESC characters removed as part of the file
name phase.
EXECUTION: Command execution is handled by the following files:
eval.c The top level routines.
redir.c Code to handle redirection of input and output.
jobs.c Code to handle forking, waiting, and job control.
exec.c Code to do path searches and the actual exec sys call.
expand.c Code to evaluate arguments.
var.c Maintains the variable symbol table. Called from expand.c.
EVAL.C: Evaltree recursively executes a parse tree. The exit
status is returned in the global variable exitstatus. The alter-
native entry evalbackcmd is called to evaluate commands in back
quotes. It saves the result in memory if the command is a buil-
tin; otherwise it forks off a child to execute the command and
connects the standard output of the child to a pipe.
JOBS.C: To create a process, you call makejob to return a job
structure, and then call forkshell (passing the job structure as
an argument) to create the process. Waitforjob waits for a job
to complete. These routines take care of process groups if job
control is defined.
REDIR.C: Ash allows file descriptors to be redirected and then
restored without forking off a child process. This is accom-
plished by duplicating the original file descriptors. The redir-
tab structure records where the file descriptors have been dupli-
cated to.
EXEC.C: The routine find_command locates a command, and enters
the command in the hash table if it is not already there. The
third argument specifies whether it is to print an error message
if the command is not found. (When a pipeline is set up,
find_command is called for all the commands in the pipeline be-
fore any forking is done, so to get the commands into the hash
table of the parent process. But to make command hashing as
transparent as possible, we silently ignore errors at that point
and only print error messages if the command cannot be found
later.)
The routine shellexec is the interface to the exec system call.
EXPAND.C: As the routine argstr generates words by parameter
expansion, command substitution and arithmetic expansion, it
performs word splitting on the result. As each word is output,
the routine expandmeta performs file name generation (if enabled).
VAR.C: Variables are stored in a hash table. Probably we should
switch to extensible hashing. The variable name is stored in the
same string as the value (using the format "name=value") so that
no string copying is needed to create the environment of a com-
mand. Variables which the shell references internally are preal-
located so that the shell can reference the values of these vari-
ables without doing a lookup.
When a program is run, the code in eval.c sticks any environment
variables which precede the command (as in "PATH=xxx command") in
the variable table as the simplest way to strip duplicates, and
then calls "environment" to get the value of the environment.
BUILTIN COMMANDS: The procedures for handling these are scat-
tered throughout the code, depending on which location appears
most appropriate. They can be recognized because their names al-
ways end in "cmd". The mapping from names to procedures is
specified in the file builtins.def, which is processed by the
mkbuiltins command.
A builtin command is invoked with argc and argv set up like a
normal program. A builtin command is allowed to overwrite its
arguments. Builtin routines can call nextopt to do option pars-
ing. This is kind of like getopt, but you don't pass argc and
argv to it. Builtin routines can also call error. This routine
normally terminates the shell (or returns to the main command
loop if the shell is interactive), but when called from a non-
special builtin command it causes the builtin command to
terminate with an exit status of 2.
The directory bltins contains commands which can be compiled in-
dependently but can also be built into the shell for efficiency
reasons. The header file bltin.h takes care of most of the
differences between the ash and the stand-alone environment.
The user should call the main routine "main", and #define main to
be the name of the routine to use when the program is linked into
ash. This #define should appear before bltin.h is included;
bltin.h will #undef main if the program is to be compiled
stand-alone. A similar approach is used for a few utilities from
bin and usr.bin.
CD.C: This file defines the cd and pwd builtins.
SIGNALS: Trap.c implements the trap command. The routine set-
signal figures out what action should be taken when a signal is
received and invokes the signal system call to set the signal ac-
tion appropriately. When a signal that a user has set a trap for
is caught, the routine "onsig" sets a flag. The routine dotrap
is called at appropriate points to actually handle the signal.
When an interrupt is caught and no trap has been set for that
signal, the routine "onint" in error.c is called.
OUTPUT: Ash uses its own output routines. There are three out-
put structures allocated. "Output" represents the standard out-
put, "errout" the standard error, and "memout" contains output
which is to be stored in memory. This last is used when a buil-
tin command appears in backquotes, to allow its output to be col-
lected without doing any I/O through the UNIX operating system.
The variables out1 and out2 normally point to output and errout,
respectively, but they are set to point to memout when appropri-
ate inside backquotes.
INPUT: The basic input routine is pgetc, which reads from the
current input file. There is a stack of input files; the current
input file is the top file on this stack. The code allows the
input to come from a string rather than a file. (This is for the
-c option and the "." and eval builtin commands.) The global
variable plinno is saved and restored when files are pushed and
popped from the stack. The parser routines store the number of
the current line in this variable.
DEBUGGING: If DEBUG is defined in shell.h, then the shell will
write debugging information to the file $HOME/trace. Most of
this is done using the TRACE macro, which takes a set of printf
arguments inside two sets of parenthesis. Example:
"TRACE(("n=%d0, n))". The double parenthesis are necessary be-
cause the preprocessor can't handle functions with a variable
number of arguments. Defining DEBUG also causes the shell to
generate a core dump if it is sent a quit signal. The tracing
code is in show.c.